Search

Hemodialysis is a membrane-based process in which solute transport from the patient’s blood to a rinsing solution (dialysate) occurs by diffusion and ultrafiltration. Devices used in hemodialysis are cylindrical modules filled with hollow-fiber membranes which allow the removal of toxic substances and metabolic wastes from the blood, but inhibit the passage of proteins and cells to the dialysate. A predictive porous-media model of hemodialysis was developed and validated against experimental data. Unlike previous literature models, it requires only basic membrane properties (hydraulic and diffusive permeabilities and reflection coefficients) instead of relying on empirically adjusted global mass transfer coefficients. The necessary porous-media characteristics, notably Darcy permeabilities and shell-side mass transfer coefficients, were obtained by combining theoretical results, CFD predictions for regular fiber arrays and experimental data for commercial modules. A parametric analysis was conducted to assess the influence of different physical quantities and operating conditions. Simulation results for different solutes showed that clearance is affected, in decreasing order of significance, by the membrane’s diffusive permeability (kM), the dialysate flow rate (QD) and the ultrafiltration flow rate (QUF). Doubling kM yields an enhancement in clearance of ~7% for urea and ~20% for B12 vitamin, while halving kM yields a decrease of ~14% for urea and ~32% for B12 vitamin. Changes in the flow rates affect clearance to a lesser extent: a 50% increase of QD yields a clearance increase of ~4−5%, while a ±10 mL/min variation of QUF with respect to the reference value of 10 mL/min leads to a clearance change of ±1.3% for urea and ±3.2% for B12 vitamin. The oncotic pressure in blood has almost no influence. Therefore, possible performance improvements, at least in terms of clearance, strictly rely on the development of a novel generation of membranes characterized by a significantly higher solute diffusive permeability.

Background Electro-membrane processes are gaining great interest in the field of desalination and brine valorisation. However, limiting current phenomena can be a bottleneck for their techno-economic performances. In the present work, the in-out distribution of current density is measured to elucidate the achievement of limiting conditions in real stacks. Materials and Methods A 10-cell pairs Electrodialysis stack (10×40 cm2 active area), equipped with four-segmented electrodes, was tested. NaCl solutions at an inlet concentration ranging from 0.5 to 60 g/l were fed at velocities of either 2 or 4 cm/s in parallel flow. Current density-voltage curves were built by applying equal increasing steps of voltage to each electrode. Outlet concentrations and current efficiency were investigated [1]. Results Figure 1 shows the current density-voltage curves for two couples of inlet concentration. Between the final tract of the ohmic region and the plateau region of the overall stack curve, the current density distribution at the four segments changes markedly. In fact, while at the first electrode the current density continues to increase, at the other three it reaches a maximum and decreases. Thus, as the voltage increases, the current concentrates in a shorter tract of the channel, while it reduces in the remaining part, becoming ineffective for desalination, due to its high resistance. This is caused by the high desalination rate in the first few centimetres, making the dilute conductivity much lower. Moreover, the longer tract of channels at high salinity gradient in the final part of the stack promotes larger diffusion, lowering the current efficiency [2]. Figure 1. Current density-voltage curves for tests at a) Cdil,IN=0.5 g/l and Cconc,IN=30 g/l and b) Cdil,IN=Cconc,IN=1 g/l. Conclusions The attainment of limiting conditions in electrodialysis stacks is strongly related to ohmic phenomena and to the distribution of current density, highlighting its importance in the design of efficient electro-membrane systems. Acknowledgments This work was supported by the SEARcularMINE (Circular Processing of Seawater Brines from Saltworks for Recovery of Valuable Raw Materials) project – Horizon 2020 programme, Grant Agreement no. 869467. The authors are grateful to REDstack B.V. and Fujifilm Manufacturing Europe B.V. for supplying stack and membranes, respectively.

Objectives The study aims to predict 3-D flow and solute concentrations fields both for blood and dialysate and overall performance parameters (such as dialysate pressure drop and clearance) for different hollow-fibre haemodialysis modules. Methods A multiscale approach was used. At small (unit cell)-scale, dialysate flow and mass transfer around straight cylindrical fibres arranged in regular lattices were simulated. At module-scale, hydraulic permeabilities and mass transfer coefficients derived from small-scale simulations were used to define two different porous media representative of blood and dialysate, sharing the same volume and exchanging solute. Simulations involved different module configurations, sharing the same membrane area but differing in geometry and inlet-outlet arrangement. Literature values of solute diffusive permeability for commercial PES membranes were used. Results and discussions For given blood and dialysate flow rates (300 and 450 mL/min, respectively), different module configurations yielded very different dialysate-side flow patterns and pressure drops, moderately different Sherwood numbers but similar values of the clearance. For urea, this ranged between 209 and 235 mL/min in most configurations, indicating that almost all the membrane area was active in most configurations. For a typical commercial geometry, the predicted clearances at different blood flow rates exhibited a satisfactory agreement with experimental measurements (maximum discrepancy ˂2%). Conclusions Using commercial membranes properties, dialysate-side mass transfer resistance is small compared with blood-side and membrane contributions. As membranes with higher diffusive permeability will be developed, the influence of dialysate-side resistance will become larger and an accurate prediction of the dialysate flow field will become more important.

Reverse osmosis (RO) is the most widespread technology to produce drinking water from seawater (SW). However, the integration of different membrane processes offers interesting alternatives. In this work, electromembrane processes were integrated with RO to desalinate real seawater in a pilot plant with 25 m3/day capacity. Electrodialysis (ED, either two-stage or single stage), shortcut reverse electrodialysis (scRED) and assisted reverse electrodialysis (ARED) pre-desalinated seawater before RO with the ED-ED-RO, ED-RO, and scRED-ARED-RO process schemes. Treated wastewater was used as salt sink in the scRED-ARED tests. The performance of the pilot plant can be summarized as follows: water recovery of ~27–51%, productivity of ~7–14 L/(m2 h) in the electromembrane processes and of ⁓19–31 L/(m2 h) in the RO process, energy consumption of 3.5–8.4 kWh/m3. The ED-RO configuration yielded the maximum productivity of the electromembrane step, while the scRED-ARED-RO integration reached the minimum energy consumption. Overall, the energy performance of the pilot plant (especially in the ED-RO and scRED-ARED-RO schemes) was comparable to that of a standalone SWRO system. The field tests demonstrated that the coupling of electromembrane processes with RO is feasible and suggest the possibility to develop alternative and competitive industrial pants for seawater desalination.

Bipolar membrane electrodialysis (BMED) is a technology combining solute and solvent dissociation to produce chemicals. In the recent decades, it has been typically studied for the production of valuable acid and base solutions from salt streams. Although many works have been devoted to the experimental investigation of BMED, only a few efforts have focused on its mathematical modelling. In the present work, a comprehensive process model based on a multi-scale approach with distributed parameters is presented for the first time. Five models related to four different dimensional scales were fully integrated to form a comprehensive tool. The integrated model was developed by using the process simulator gPROMS Model builder and was based on a semi-empirical approach combining high prediction accuracy and low computational demand. Once validated through a wide range of experimental data, the model capability was shown by carrying out a broad sensitivity analysis assessing the performance of the BMED technology for industrial-scale applications. Results showed how the performance of a BMED unit changes with both varying process conditions and the installed membrane area. Particularly, the non-ideal phenomena that reduce the produced NaOH concentration and increase the energy consumption were thoroughly investigated. Finally, this study demonstrated that a Levelized Cost Of Caustic Soda of about 280 € ton-1NaOH can be obtained, thus making this technology a possible candidate for the industrial production of caustic soda from brines in the future.

Reverse Electrodialysis (RED) extracts electrical energy from the salinity difference between two solutions using selective ion exchange membranes. In RED, conditions yielding a large net power density (NPD) are generally desired, due to the still large cost of the membranes. NPD depends on a large number of physical and geometric parameters. Some of these, for example the inlet concentrations of concentrate and diluate, can be regarded as “scenario” variables, imposed by external constraints (e.g., availability) or chosen by different criteria than NPD maximization. Others, namely the thicknesses HCONC, HDIL and the velocities UCONC, UDIL in the concentrate and diluate channels, can be regarded as free design parameters and can be chosen so as to maximize NPD. In the present study, a simplified model of a RED stack was coupled with an optimization algorithm in order to determine the conditions of maximum NPD in the space of the variables HCONC, HDIL,UCONC, UDIL for different sets of “scenario” variables. The study shows that an optimal choice of the free design parameters for any given scenario, as opposed to the adoption of standard fixed values for the same parameters, may provide significant improvements in NPD.

The problem of mass transfer in ducts with transpiring walls is analysed: the concepts of “solvent” and “solute” fluxes are introduced, all possible sign combinations for these fluxes are considered, and relevant examples from membrane processes such as electrodialysis, reverse osmosis and filtration are identified. Besides the dimensionless numbers commonly defined in studying flow and mass transfer problems, new dimensionless quantities appropriate to transpiration problems are introduced, and their limiting values, associated with “drying”, “desalting” and “saturation” conditions, are identified. A simple model predicting the Sherwood number Sh under all possible flux sign combinations is derived from the single simplifying assumption that concentration profiles remain self-similar (so that the Sherwood number based on diffusion only remains unchanged) also under transpiration conditions. The simple model provides not only local values of Sh, but also its axial profiles. Predictions are validated against fully predictive CFD results, not based on the above simplifying assumption, and a good agreement is demonstrated provided the transpiration rate complies with certain limitations, depending on the Schmidt number.

This Special Issue of Membranes journal focuses on electromembrane processes and is motivated by the increasing interest of the scientific community towards their characterization by experiments and modelling for several applications [...]

Objectives: The model proposed aims to predict how geometric, transport and operative parameters affect the performances of hollow-fibre membrane modules for haemodialysis, especially solute clearance. Methods: A two-scale approach was used. Preliminarily, dialysate flow and mass transfer around fibre bundles were simulated at Unit Cell level, i.e. in a single periodic unit of the bundle. For a given porosity, both regular lattices (square or hexagonal) and random fibre arrangements were studied. From the predicted friction coefficients and Sherwood numbers, permeability and solute exchange terms were derived to be used in a porous media model of the whole module. Solute concentrations on the blood and dialysate sides were alternatively simulated in an iterative fashion and were coupled by appropriate sink/source terms describing solute transfer. The model predicted 3-D flow fields and solute concentrations both in the blood and in the dialysate, as well as overall performance parameters such as clearance and pressure drop, for given module configuration, solute species and membrane properties. Results and discussions: At unit cell level, the dialysate side Sherwood number depended weakly on the axial velocity, but increased greatly in the presence of cross flow (up to 2.5 times for transversal velocities of ~1/4th the axial velocity). The presence of cross flow affected little the axial permeability and thus the axial flow rate. At module level, several cylindrical configurations sharing the same membrane area but differing in aspect ratio and inlet/outlet arrangement yielded very different pressure drops but similar values of the clearance. Conclusions: For typical commercial membrane properties, most of the resistance to mass transport is located in the membrane and the flow field has only a minor influence. Improving dialysate-side mass transfer coefficients and flow distribution can have beneficial effects on clearance only if membrane performances are also upgraded.

Flow and heat or mass transfer in channels provided with woven spacers made up of mutually orthogonal filaments were studied by Computational Fluid Dynamics. The problem addressed was the combined effect of the parameters that characterize the process: pitch to height ratio P/H (2, 3 and 4), flow attack angle θ (0, 7, 15, 20, 30, 40 and 45°) and Reynolds number Re (from ~1 to ~4000). The Prandtl number was 4.33, representative of water at ~40°C, while the Schmidt number was 600, representative of NaCl solutions. Simulations were performed by the finite volume code Ansys CFX™ 18.1 using very fine grids of ~6 to ~14 million volumes. For Re > ~400, the SST turbulence model was used to predict flow and heat transfer; no simulations of mass transfer were performed in the turbulent regime because the resolution of the diffusive sublayer would have required a prohibitive number of grid points. Results were validated against experimental data, including results obtained by Liquid Crystal Thermography and Digital Image Processing. The flow attack angle θ = 45° was the most effective for mixing (higher Nusselt number, Nu, and Sherwood number, Sh) and caused lower values of friction coefficient (f). In the range investigated, increasing the pitch to height ratio P/H caused Nu, Sh and f all to decrease. Therefore, the highest values of Sh and Nu were provided by the configuration P/H = 2, θ = 45°. Compared with non-woven spacers, woven spacers provided a better mixing, especially at intermediate values of Re, but at the expenses of higher pressure drops.

Numerical simulations were conducted for fully developed, steady-state flow with mass transfer in fiber bundles arranged in regular lattices. The porosity was 0.5 and the Schmidt number 500. Several combinations of axial flow, transverse flow and flow attack angles in the cross-section plane were considered. The axial and transverse Reynolds numbers Rez , ReT were made to vary from 10(^−4) to 10(^2). Concentration boundary conditions, and the definition of an average Sherwood number, were addressed. Results for the hydraulic permeability were compared with the literature. Both hexagonal and square lattices were found to be hydraulically almost isotropic up to transverse flow Reynolds numbers ReT of ~10, while they behaved anisotropically in regard to mass transfer even at ReT as low as 0.1. A larger anisotropy was exhibited by the square lattice. In mixed (axial + transverse) flow, the transverse friction coefficient was almost completely unaffected by the simultaneous presence of axial flow, while the axial friction coefficient (and thus the axial pressure loss) increased with the transverse Reynolds number for ReT > ~5-10. In regard to mass transfer, the Sherwood number settled in all cases to the higher between the Sherwood number in purely transverse flow and that in purely axial flow.

This paper presents a comprehensive review of studies on electrodialysis (ED) applications in wastewater treatment, outlining the current status and the future prospect. ED is a membrane process of separation under the action of an electric field, where ions are selectively transported across ion-exchange membranes. ED of both conventional or unconventional fashion has been tested to treat several waste or spent aqueous solutions, including effluents from various industrial processes, municipal wastewater or salt water treatment plants, and animal farms. Properties such as selectivity, high separation efficiency, and chemical-free treatment make ED methods adequate for desalination and other treatments with significant environmental benefits. ED technologies can be used in operations of concentration, dilution, desalination, regeneration, and valorisation to reclaim wastewater and recover water and/or other products, e.g., heavy metal ions, salts, acids/bases, nutrients, and organics, or electrical energy. Intense research activity has been directed towards developing enhanced or novel systems, showing that zero or minimal liquid discharge approaches can be techno-economically affordable and competitive. Despite few real plants having been installed, recent developments are opening new routes for the large-scale use of ED techniques in a plethora of treatment processes for wastewater.

The theoretical energy density extractable from acidic and alkaline solutions is higher than 20 kWh m−3 of single solution when mixing 1 M concentrated streams. Therefore, acidic and alkaline industrial wastewater have a huge potential for the recovery of energy. To this purpose, bipolar membrane reverse electrodialysis (BMRED) is an interesting, yet poorly studied technology for the conversion of the mixing entropy of solutions at different pH into electricity. Although it shows promising performance, only few works have been presented in the literature so far, and no comprehensive models have been developed yet. This work presents a mathematical multi-scale model based on a semi-empirical approach. The model was validated against experimental data and was applied over a variety of operating conditions, showing that it may represent an effective tool for the prediction of the BMRED performance. A sensitivity analysis was performed in two different scenarios, i.e. (i) a reference case and (ii) an improved case with high-performance membrane properties. A Net Power Density of ~15 W m−2 was predicted in the reference scenario with 1 M HCl and NaOH solutions, but it increased significantly by simulating high-performance membranes. A simulated scheme for an industrial application yielded an energy density of ~50 kWh m−3 (of acid solution) with an energy efficiency of ~80–90% in the improved scenario.

This work presents computational fluid dynamics simulations aimed at characterizing flow and mass/heat transport mechanisms in spacer-filled channels for membrane processes, with particular reference to (reverse) electrodyalisis and membrane distillation.

La Acid/Base Flow Battery (AB-FB) è una batteria innovativa che rappresenta un metodo sostenibile di immagazzinare energia elettrica. Questa batteria può, in linea teorica, garantire una densità di energia pari a 11 kWh/m3 che è un valore maggiore rispetto a quello fornito dalle tecnologie attualmente più utilizzate nel mondo quali quelle che sfruttano l’energia idroelettrica, l’energia osmotica o ancora l’aria compressa. La batteria AB-FB accumula energia sottoforma di gradienti di salinità e di pH. In particolare, essa funziona grazie ad una unità chiamata “stack” al cui interno ospita una serie di canali che permettono il passaggio di soluzioni elettrolitiche a diversi pH e concentrazioni delle specie chimiche. Questi canali sono separati da membrane monopolari e bipolari a scambio ionico. Nello stack sono coinvolti due processi a membrana: l’elettrodialisi con membrane bipolari (ED-BM) e il suo processo inverso ovvero la elettrodialisi inversa con membrane bipolari (RED-BM). Il primo avviene durante la carica e il secondo durante la scarica della batteria. Questo lavoro di ricerca mira a predire le performance di questo dispositivo attraverso la messa a punto di un modello matematico basato su un approccio di tipo multi-scala. Nello specifico, il singolo canale dello stack rappresenta la scala più bassa del modello, nel quale simulazioni di fluidodinamica computazionale permettono di predire le perdite di carico e il fenomeno della polarizzazione di concentrazione lungo il canale. Il livello medio-basso del modello riguarda il cosiddetto “triplet” ossia l’unità ripetitiva dello stack. A questa scala dimensionale vengono calcolati tutti i flussi attraverso la membrana cioè quello migrativo, diffusivo, osmotico ed elettro-osmotico. Alla scala medio-alta, questo modello consente di predire le perdite di carico e le correnti parassite nei collettori e nei distributori dello stack. In particolare viene calcolata l’incidenza delle correnti di corto-circuito ionico (correnti parassite). Infine, la scala più alta del modello include il calcolo delle equazioni di bilancio di materia all’interno dei recipienti esterni e le perdite di carico nei tubi di collegamento dello stack ai recipienti stessi. Tutte le scale del modello sono completamente integrate l’un l’altra in un unico e originale strumento di simulazione che è in grado di predire parametri di performance della batteria come la densità di potenza prodotta oppure il valore di Round Trip Efficiency. Un’analisi di sensitività è stata effettuata variando caratteristiche geometriche come il diametro del distributore e del collettore e condizioni operative come la densità di corrente durante la carica e la scarica in un ampio range di valori. I risultati ottenuti mostrano che le correnti ioniche parassite attraverso i collettori e i distributori possono rappresentare il limite più importante per questa tecnologia. La progettazione della batteria con geometrie e condizioni operative ottimali possono portare alla riduzione dei fenomeni parassiti e a valori di Round Trip Efficiency anche maggiori del 65% rendendo così questa batteria competitiva rispetto alle tecnologie attualmente più utilizzate.

The Acid/Base Flow Battery (AB-FB) is an innovative and sustainable way to store electric energy. It can theoretically guarantee an energy density of about 11 kWh/m3, which is higher than that provided by pumped hydropower, osmotic energy storage and compressed air. The AB-FB stores energy as pH and salinity gradients by employing a stack provided with (i) channels, hosting the solutions at difference pH and concentrations, separated by (ii) monopolar and bipolar ion exchange membranes. Two different membrane processes are involved: the Bipolar Membrane Electrodialysis (ED-BM) as charging step and its opposite, Bipolar Membrane Reverse Electrodialysis (RED-BM) as discharging step. The present work aims at predicting the performance of this AB-FB energy storage device via the development of a mathematical model based on a multi-scale approach. The channel represents the lowest scale of the model where Computational Fluid Dynamic Simulations are adopted to predict pressure drops and polarization phenomena. The middle-low-scale model concerns the triplet (which is the repeating unit of the stack) where all membrane fluxes (i.e. ohmic, diffusive, osmotic, etc.) are calculated. The middle-high-scale is devoted to predicting the pressure losses and the ionic losses (due to ionic short-circuit currents) in the manifolds. The highest scale includes all equations relevant to the connections of the stack with the tanks used to store the solutions. All model-scales are fully integrated, thus representing an original tool able to predict the flow battery performance parameters such as the power density produced and the overall battery round trip efficiency. A sensitivity analysis is performed by letting geometrical features (e.g. manifold size) and operating conditions (e.g. charge and discharge current density) to vary in a wide range of values. Main results show that the parasitic currents in the manifolds may represent the main limit to the present technology among all the detrimental phenomena. Suitable geometries and operating conditions can be adopted to reduce their effect thus leading to round-trip efficiencies higher than 65%.

In the present work, we experimenrtally investigated the effect of different membrane profiles on the LCD, testing also different operating conditions.

L’elettrodialisi (ED) è una promettente tecnologia a membrana utilizzata in diversi campi, ad esempio nella dissalazione delle acque e nell’industria alimentare. L’ED usa un potenziale elettrico per indurre una migrazione selettiva di cationi ed anioni da una soluzione elettrolitica ad un’altra, sfruttando membrane a scambio ionico. Membrane anioniche e cationiche sono alternativamente collocate all’interno di una unità ED. A queste sono solitamente interposti spaziatori che prevengono il contatto tra le membrane e delineano i canali in cui scorrono le soluzioni. L’utilizzo di membrane profilate consente di costruire unità prive di tradizionali spaziatori a rete non conduttivi. In genere, la differenza di pressione tra le due soluzioni (trans-membrane pressure, TMP) in unità ED assume valori modesti o trascurabili (TMP

Reverse electrodialysis (RED) is a blue energy technology for clean and sustainable electricity harvesting from the mixing entropy of salinity gradients. Recently, many efforts have been devoted to improving the performance of RED units by developing new ion-exchange membranes and by reducing the detrimental phenomena affecting the process. Among these sources of “irreversibility”, the shortcut currents (or parasitic currents) flowing through alternative pathways may affect the process efficiency. Although such phenomena occur in several electrochemical processes (e.g. fuel cells, bipolar plate cells and vanadium redox flow batteries), they have received a poor attention in RED units. In this work, a process simulator with distributed parameters was developed and experimentally validated to characterize the shortcut currents and to assess their impact in RED stack performance under different designs and operating conditions. Results showed that shortcut currents can play a crucial role in stacks with a large number of cell pairs when the electrical resistance of the parasitic pathways is relatively low, e.g. configurations with concentrated brines, high resistance membranes, short channels or large manifolds. Future designs of efficient industrial-scale units cannot ignore these aspects. Finally, the model can be easily adapted for the simulation of electrodialysis and other electromembrane processes.

Selectivity, high recovery, and chemical-free operation are strengths of electrodialysis. Different configurations have been proposed for several applications. Effluents from municipal wastewater treatment plants (including sludge and supernatants), desalination plants, and animal farms can be treated for recovering water, nutrients, salts, and acids/bases. Although many applications are technoeconomically feasible and competitive with other zero liquid discharge systems, only a few real plants have been installed. However, the research is currently very active, thus paving the way for a widespread use at large scale in the next future.

Electrodialysis and related processes have huge potential in the treatment of effluents from a variety of industrial processes. They can recover water and other valuable products, including heavy metal ions, acids and bases, nutrients, and organics. In recent years, novel and improved systems have been continuously developed as a result of research in the field, showing that the (near) zero liquid discharge approach can be affordable in several industrial applications. A larger market share is expected in the near future.

Bipolar Membrane Reverse Electrodialysis (BMRED) can be used to produce electricity exploiting acid-base neutralization, thus representing a valuable route in reusing waste streams. The present work investigates the performance of a lab-scale BMRED module under several operating conditions. By feeding the stack with 1 M HCl and NaOH streams, a maximum power density of ~17 W m−2 was obtained at 100 A m−2 with a 10-triplet stack with a flow velocity of 1 cm s−1, while an energy density of ~10 kWh m−3 acid could be extracted by a complete neutralization. Parasitic currents along feed and drain manifolds significantly affected the performance of the stack when equipped with a higher number of triplets. The apparent permselectivity at 1 M acid and base decreased from 93% with the five-triplet stack to 54% with the 38-triplet stack, which exhibited lower values (~35% less) of power density. An important role may be played also by the presence of NaCl in the acidic and alkaline solutions. With a low number of triplets, the added salt had almost negligible effects. However, with a higher number of triplets it led to a reduction of 23.4–45.7% in power density. The risk of membrane delamination is another aspect that can limit the process performance. However, overall, the present results highlight the high potential of BMRED systems as a productive way of neutralizing waste solutions for energy harvesting.

Electrodialysis (ED) has recently gained much attention in the wide field of desalination and water treatment. However, energy consumption and capital costs may impair the process competitiveness. In this regard, limiting current density (LCD) and current efficiency (η) are key performance parameters for optimized ED systems. In this work, an experimental campaign was carried out characterizing the performance of ED stacks when adopting asymmetrically profiled membranes. Current–voltage curves were recorded under different operating conditions mimicking the operation of brackish water or seawater desalination units. Results showed that there was a preferable direction of the electric current relative to the membrane profiles, which provided higher values of LCD and of maximum η. Stacks with Overlapped Crossed Filaments profiled membranes performed better than conventional ED stacks with flat membranes and spacers by increasing the LCD (by ~20% under various operating conditions) and the maximum η (e.g., from ~55–65% to ~70–73% at 0.5 – 60 g/l inlet concentrations of diluate and brine, respectively). The specific energy consumption was significantly reduced (even more than 50%). On the contrary, the investigated pillar-profiled membranes exhibited the worst performances. The present results suggest that well-designed profiled membranes can reduce the costs of desalination via ED.

Membrane-based processes have gained a relevant role in many engineering applications. Much effort has been devoted to thoroughly understand the fundamental phenomena behind them. However, membrane deformation has been taken into consideration only recently, although much evidence has shown its impacts in many applications. This work presents a novel 2-D, multi-scale, semi-empirical process model able to predict the behavior and the performance of Electrodialysis (ED) systems in cross-flow configurations in the presence and absence of local membrane deformations. The model exploits the results and the simulation approaches of previous fluid-structure investigations performed by the authors. Low-scale numerical simulations are coupled with a high-scale model to predict the redistribution of channel height, flow rate, friction coefficient and Sherwood number in ED stacks caused by local membrane deformations. Finally, salt and water fluxes, mass balances and electrochemical quantities are computed to assess the performances of cross-flow ED stacks. Different test cases have been simulated for the desalination of seawater by two-stage ED. Interestingly, membrane deformation is found to reduce, albeit slightly, the energy consumption. More pronounced effects are expected if thinner or less stiff membranes are used.

In the present work, in order to explore such issues on the LCD identification, we performed in-situ measurements with ED units, assessing the influence of operating conditions and validating a purposely implemented process simulator, which was then used for further investigation

The present work describes the development of the first dynamic model for CED

In this work, we developed an advanced model useful for the design of profiled IEMs, based on the coupled simulation of local mechanical deformations and of fluid dynamics and associated mass transport phenomena within deformed channels

The need to reduce energy consumption in seawater reverse osmosis processes has pushed research towards the development of new hybrid systems in which, for example, other membrane processes can be used to pre-treat seawater. Electrodialysis and reverse electrodialysis can act as a dilution step before seawater enters the RO unit, thus leading to an important energy saving in RO. In this work, two coupled models are proposed for the RED-RO and ED-RO system. Each process model was validated before being used for a sensitivity analysis in which the effect of the integration on the cost saving in the overall process was assessed. The analysis was performed by changing (R)ED voltage and RO pressure and considering three different scenarios: a standard scenario, an optimist scenario with a lower membrane cost and a pessimistic scenario with a lower electricity cost and comparing the result with the a standalone RO process. Negative values of “Cost Saving” were found when an excessive dilution step was performed before the RO, while competitive scenarios were found by optimizing the dilution extent, especially for the RED-RO case.

In recent years, water and energy supply issues have boosted a noticeable interest in the scientific community on electromembrane processes such as electrodialysis and reverse electrodialysis. In order to gain an important place in the industrial market, technological challenges on various aspects are involved for the optimization of these processes. In this context, profiled membranes exhibit interesting performances and offer countless geometric alternatives. However, the mechanical behavior of the membranes and its interaction with fluid dynamics has been poorly investigated so far. In membrane-based processes, a trans-membrane pressure (Ptm) between the different solutions flowing through a module may be a design feature or may arise for various reasons, including flow arrangement and differences in physical properties, flow rate or friction coefficient. This leads to local deformations of membranes and channels, affecting flow and mass transfer characteristics, thus causing uneven distributions of flow and mass fluxes, which worsen the process performance. In this work, we developed an integrated model for the numerical simulation of local mechanical deformations and of fluid dynamics and associated mass transport phenomena inside deformed channels. Two diverse profiled membrane types (“overlapped cross filaments”, OCF, and “round pillars”, RP) were simulated under conditions representative of (reverse) electrodialysis and under the assumption of perfectly elastic behaviour. 3-D simulations of a couple of membranes and of the interposed fluid were conducted by the unit cell approach (periodic domain). The Ansys Mechanical 18 (Workbench) and the Ansys CFX 18 software was used. The selected geometries were simulated under Ptm ranging from -0.4 to +0.4 bar, computing expanded and compressed configurations. Then, CFD simulations of the deformed channels were performed, showing significant effects of the deformation on fluid flow and mass transfer. The influence of Ptm was to increase friction under compression conditions (up to ∼2.2-2.5 times) and to reduce it under expansion conditions (but to a lesser extent, i.e. up to ∼50-60%). Overall, compression enhanced mass transfer and expansion reduced it, but with smaller and more complex effects than on friction. The influence of the flow attack angle was negligible for friction, but more significant for mass transfer. In future works the same simulation approach will be adopted in order to compute also the Ohmic resistance in deformed configurations. The simulation results will be implemented in the form of correlations into higher-scale models, in order to study distributions of flow, mass transfer and Ohmic resistance in whole channels. The method proposed can be extended to other membrane applications with minor modifications.

The future energetic supply based on the massive use of renewable sources poses issues linked to fluctuations of power produced and consumed, thus requiring the use of energy storage systems.

Reverse Electrodialysis (RED) harvests electrical energy from a salinity gradient. The maximum obtainable net power density (NPD) depends on many physical and geometric variables. Some have a monotonic (beneficial or detrimental) influence on NPD, and can be regarded as “scenario” variables chosen by criteria other than NPD maximization. Others, namely the thicknesses HCONC, HDIL and the velocities UCONC, UDIL in the concentrate and diluate channels, have contrasting effects, so that the NPD maximum is obtained for some intermediate values of these parameters. A 1-D model of a RED stack was coupled here with an optimization algorithm to determine the conditions of maximum NPD in the space of the variables HCONC, HDIL,UCONC, UDIL for different combinations of the “scenario” variables. The model accounts for entrance effects, property variation, concentration polarization, axial concentration changes, osmotic, electro-osmotic and diffusive fluxes, and can deal with complex channel geometries using Ohmic resistances, friction factors and mass transfer coefficients computed by 3-D simulations.

A new hierarchical model for the electrodialysis (ED) process is presented. The model has been implemented into gPROMs Modelbuilder (PSE), allowing the development of a distributed-parameters simulation tool that combines the effectiveness of a semi-empirical modelling approach to the flexibility of a layered arrangement of modelling scales. Thanks to its structure, the tool makes possible the simulation of many different and complex layouts, requiring only membrane properties as input parameters (e.g. membrane resistance or salt and water permeability). The model has been validated against original experimental data obtained from a lab scale ED test rig. Simulation results concerning a 4-stage treatment of seawater and dynamic batch operations of brackish water desalination are presented, showing how the model can be effectively used for predictive purposes and for providing useful insights on design and optimisation.